config.json

{
  "architectures": [
    "FastSpeech2ConformerWithHifiGan"
  ],
  "dtype": "float32",
  "model_config": {
    "convolution_bias": true,
    "decoder_attention_dropout_rate": 0.2,
    "decoder_concat_after": false,
    "decoder_config": {
      "attention_dropout_rate": 0.2,
      "concat_after": false,
      "dropout_rate": 0.2,
      "kernel_size": 31,
      "layers": 1,
      "linear_units": 384,
      "normalize_before": false,
      "num_attention_heads": 2,
      "positional_dropout_rate": 0.2
    },
    "decoder_dropout_rate": 0.2,
    "decoder_kernel_size": 31,
    "decoder_layers": 1,
    "decoder_linear_units": 384,
    "decoder_normalize_before": false,
    "decoder_num_attention_heads": 2,
    "decoder_positional_dropout_rate": 0.2,
    "duration_predictor_channels": 256,
    "duration_predictor_dropout_rate": 0.2,
    "duration_predictor_kernel_size": 3,
    "duration_predictor_layers": 1,
    "encoder_attention_dropout_rate": 0.2,
    "encoder_concat_after": false,
    "encoder_config": {
      "attention_dropout_rate": 0.2,
      "concat_after": false,
      "dropout_rate": 0.2,
      "kernel_size": 7,
      "layers": 1,
      "linear_units": 384,
      "normalize_before": false,
      "num_attention_heads": 2,
      "positional_dropout_rate": 0.2
    },
    "encoder_dropout_rate": 0.2,
    "encoder_kernel_size": 7,
    "encoder_layers": 1,
    "encoder_linear_units": 384,
    "encoder_normalize_before": false,
    "encoder_num_attention_heads": 2,
    "encoder_positional_dropout_rate": 0.2,
    "energy_embed_dropout": 0.0,
    "energy_embed_kernel_size": 1,
    "energy_predictor_channels": 256,
    "energy_predictor_dropout": 0.5,
    "energy_predictor_kernel_size": 3,
    "energy_predictor_layers": 1,
    "hidden_size": 24,
    "max_source_positions": 5000,
    "model_type": "fastspeech2_conformer",
    "num_languages": null,
    "num_mel_bins": 8,
    "num_speakers": null,
    "pitch_embed_dropout": 0.0,
    "pitch_embed_kernel_size": 1,
    "pitch_predictor_channels": 256,
    "pitch_predictor_dropout": 0.5,
    "pitch_predictor_kernel_size": 5,
    "pitch_predictor_layers": 1,
    "positionwise_conv_kernel_size": 3,
    "reduction_factor": 1,
    "speaker_embed_dim": null,
    "speaking_speed": 1.0,
    "speech_decoder_postnet_dropout": 0.5,
    "speech_decoder_postnet_kernel": 5,
    "speech_decoder_postnet_layers": 2,
    "speech_decoder_postnet_units": 128,
    "stop_gradient_from_energy_predictor": false,
    "stop_gradient_from_pitch_predictor": true,
    "use_cnn_in_conformer": true,
    "use_macaron_style_in_conformer": true,
    "use_masking": true,
    "use_weighted_masking": false,
    "vocab_size": 78
  },
  "model_type": "fastspeech2_conformer_with_hifigan",
  "transformers_version": "5.16.0.dev0",
  "vocoder_config": {
    "initializer_range": 0.01,
    "leaky_relu_slope": 0.1,
    "model_in_dim": 8,
    "model_type": "fastspeech2_conformer_hifigan",
    "normalize_before": true,
    "resblock_dilation_sizes": [
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ]
    ],
    "resblock_kernel_sizes": [
      3,
      7,
      11
    ],
    "upsample_initial_channel": 64,
    "upsample_kernel_sizes": [
      16,
      16,
      4,
      4
    ],
    "upsample_rates": [
      8,
      8,
      2,
      2
    ]
  }
}